Purpose: To ascertain the dose–toxicity relationship for the prevalence of self-reported trismus in long-term survivors after intensity-modulated radiotherapy (IMRT) for oropharyngeal carcinoma (OPC). Materials and methods: Self-reported mouth opening was ascertained prospectively via a cross-sectional survey of OPC survivors using the intraoral finger-test. RT dose–volume histograms (DVHs) were generated for the following masticatory regions of interest: medial pterygoid, lateral pterygoid, and masseter muscles which were designated as ipsilateral or contralateral to the primary tumor. Trismus was defined as self-reported mouth opening of <3 finger-widths. Recursive partitioning analysis (RPA) was performed to identify the dose–volume thresholds associated with late trismus. Results: At a median follow-up time of 72 months (95% CI 68–74), 168 of the 587 (29%) survey respondents reported late trismus. Multivariate analysis demonstrated a significant association between late trismus and the following clinical variables: tonsillar primary site, advanced T stage, or higher total RT dose. RPA showed DVH-derived ipsilateral lateral pterygoid (ILP) mean dose of 61 Gy and volume receiving 27 Gy of at least 98.6% were independently associated with late trismus. The association between the ILP dosimetric parameters and the prevalence of late trismus was maintained after adjustment for clinical variables. Conclusion: The integral dose of IMRT results in unavoidable low/intermediate dose to non-target masticatory muscles that is associated with increased prevalence of late trismus in OPC survivors. Whenever clinically and technically applicable, applying the proposed dosimetric constraints to the ILP (V27 <98.6 and Dmean <61 Gy) may reduce the prevalence of late trismus after IMRT for OPC patients.